212 BULLETIN OF THE BUREAU OF FISHERIES. 
of the imagos, however, consists chiefly of soft-bodied insects, destitute of elytra, or 
of insects from which the harder parts, if present, are carefully rejected. Consequently 
the only things that can be recognized with any certainty in the contents of the imago’s 
digestive canal are an occasional mandible or maxilla, portions of the wings of various 
insects, legs and antenne, scales of butterflies and moths, hairs, and claws. The 
accurate identification of such minutie is a painstaking and laborious task, and the 
frequency with which the species or even the genus must be left undetermined is not 
surprising. 
Again, the nymph swallows much of its food whole or at least in large fragments, so 
that the relation of the various parts remains undisturbed. The imago, on the other 
hand, believes in thoroughly masticating its food, and every mouthful is chewed into 
fine fragments before being swallowed. At the same time such parts as the wings and 
legs, which would be of great value for identification, and even the harder tissues of 
the body, are carefully rejected. Occasionally fragments of a wing or elytron are 
sometimes included, but they are usually badly torn and often lack the very part that 
is wanted. The imago is particularly fond of teneral insects, whose chitin has not yet 
hardened, and whose pigment markings have not been developed. Such insects, after 
being chewed and swallowed, form an indistinguishable mass in which there is very little 
hope of finding anything that can be identified. 
A third difficulty is found in the fact that, although the digestion of the nymph’s food 
is comparatively slow and the large fragments are recognizable for some time after they 
have been swallowed, the food of the imago, on the other hand, is digested with excep- 
tional rapidity and must be examined as soon as it is swallowed, if anything definite is to 
be hoped for. Even the short space of time between the insertion of the insect into a 
cyanide bottle and its subsequent death is sufficient to materially affect the contents of 
the alimentary canal, and the changes apparently continue a short time after the insect’s 
death. To obviate this, good results were obtained by making an incision in the thorax 
and abdomen, and then plunging the imago into 95 per cent alcohol as soon as it was 
taken from the net. All the examinations of the alimentary canal here tabulated were 
made in this way. 
In view of these difficulties the most feasible method of determining the food of the 
imago is to watch it while feeding and capture it with enough of its food still uneaten to 
render identification possible. That this method has proved very satisfactory is 
shown by the frequency with which it appears in the following statements. In addition, 
all the available American records have been included, with acknowledgment of their 
source. 
Foop oF Gompuip Imacos, Fairport, IowA, 1916. 
Gomphus fraternus: , 
Diptera—House fly, Musca domestica............2++.- Captured while eating. 
Trichoptera—Caddisfly, undetermined .....- EAN OSES. . In alimentary canal. 
Odonata— 
Erythemts simplicicollis. .0.5..0. 06. 0ce cece cee cee ee Needham and Hart, rgor, p. 64. 
Tabellula tuctuosa® 5... )satetpioe casas skh Had es Captured while eating. 
ALT GtG MUSE Petre te nist svarnict state bop Tals alc la a Do. 
‘Péeneraltdragoniies sto 2+... comes reete ete cetlac In alimentary canal, 
